The AI Economy — Research Report

A data-driven, fully source-cited research report mapping the AI economy layer by layer — from the electricity grid up through chips, models, and applications — with numbers, visuals, and 2030 projections. The standard is rigour: every figure traces to a verified source, so the work can stand up to scrutiny and inform real AI-economy analysis.

Status

Layer	State
L0 · Energy & Power	✅ Deep chapter complete & frozen (see docs/HANDOFF_L0_Energy.md)
L1 · Semiconductors	In progress (20 rows researched, awaiting sign-off)
L2–L7, predictions, regulation, futures	⏳ planned (see docs/PROJECT_CONTEXT.md)

Ledger: 75 Confirmed / 30 Proposed rows.

Repository layout

.
├── README.md                  ← you are here
├── docs/                      project context, delivery format, handoff
│   ├── PROJECT_CONTEXT.md      the report's goal, story arc, strategy
│   ├── DELIVERY_FORMAT.md      how pages are built (data-story block, rules)
│   └── HANDOFF_L0_Energy.md    onboarding for anyone continuing L0
├── blueprints/                content specs that drive each rendered page
│   ├── README.md               how the blueprint → render loop works
│   ├── deep_energy.md          spec for the deep Energy chapter
│   └── ch1_L0_energy.md        spec for the high-level Ch.1 L0 section
├── data/                      the source of truth
│   ├── AI_Economy_Data_Ledger.xlsx   every number, with source + status
│   ├── data.json              exported Confirmed rows (for the website)
│   ├── maps_b64.json           rendered choropleths (base64 PNG)
│   └── reference/country_energy_2024.csv   31-country dataset (maps)
├── scripts/                   regenerate everything (run from repo root)
│   ├── export_ledger_json.py   ledger → data.json
│   ├── make_maps.py            country data → geographic choropleths
│   └── build_deep_energy.py    assembles the deep chapter HTML
├── site/                      rendered output (open in a browser)
│   ├── chapter_energy_deep.html      the deep Energy & Power chapter
│   └── chapter1_L0_energy.html       the high-level Ch.1 L0 section
└── sources/                   background source material & WIP drafts

How it works

1. The ledger is the single source of truth. data/AI_Economy_Data_Ledger.xlsx holds one row per data point with value, source, confidence, and a Status gate. Nothing reaches a chart or the site unless its status is Confirmed.
2. The blueprint → render loop. Edit a file in blueprints/ to change what a page says/shows → research any new facts and add them as Proposed ledger rows → get sign-off (flip to Confirmed) → re-render. The HTML is disposable output; the ledger + blueprints are the durable assets.
3. Two-tier data model. Curated story numbers live in the ledger (cell-level sign-off); bulk country datasets (e.g. data/reference/) are approved at the source level.
4. Static rendering. All charts are self-contained — inline SVG or embedded PNG, no external scripts — so pages render anywhere. Geographic choropleths are produced offline (scripts/make_maps.py) and embedded as images.

Rebuild

pip install openpyxl pygal pygal_maps_world cairosvg --break-system-packages
# run from the repo root:
python scripts/export_ledger_json.py     # ledger → data/data.json
python scripts/make_maps.py              # → data/maps_b64.json (geographic maps)
python scripts/build_deep_energy.py      # → site/chapter_energy_deep.html

Where to start

New here? Read docs/HANDOFF_L0_Energy.md first (it points to everything else), then open site/chapter_energy_deep.html in a browser.